DSH Plugins Marketplace

DSH Plugins

Plugins

/

Memory & Context

/

dsh-obsidian-bridge

d

dsh-obsidian-bridge

Identified

Connect a local Obsidian vault to DeepSeek Harness: local Markdown indexing with keyword search by default and optional hybrid or semantic search through `obsidian_*` agent tools (hybrid and semantic

Obsidian Vault Studio

Obsidian Vault Studio dashboard

Connect a local Obsidian vault to DeepSeek Harness. After you grant access, the model can search your notes, read relevant passages, build grounded context, and create or append notes through a proposal-first workflow. The index, embeddings, and statistics stay on your machine.

中文

npm versionGitHub stars

What it does

  • Local index — parses Markdown, frontmatter, tags, headings, and modification times without uploading the whole vault.
  • Practical retrieval — keyword search by default, with English prefix matching and live results; optional OpenAI-compatible or Ollama embeddings enable hybrid and semantic modes.
  • Conversation access switch — model access is off by default. When off, every obsidian_* Agent tool is denied; your own search and note browsing keep working.
  • Safe writes — the model can only propose a create or append. Nothing is written until you approve; writes are backed up and can be undone. There are no delete, move, rename, or overwrite tools.
  • Dashboard — note, word, tag, and chunk counts; a daily heatmap; a 14-day activity curve; Agent call counters; recent notes and tag distribution.
  • Local by design — index, vectors, settings, and usage counters all stay on your machine.

Install

Requirements

  • DeepSeek Harness 0.2.x desktop, with third-party plugin installation enabled
  • A local Obsidian vault

Install source

DeepSeek Harness can install the plugin from either a GitHub repository or an npm package:

  • GitHub repository: https://github.com/spark-ai-boy/dsh-obsidian-bridge
  • npm package: dsh-obsidian-bridge

Steps

  1. Open Plugins in the DeepSeek Harness sidebar.
  2. Choose Add plugin → Install third-party plugin.
  3. Enter either the GitHub repository URL or the npm package name above, install it, then click Enable.
  4. Open Obsidian Vault Studio in the sidebar.

Obsidian Vault Studio installation

First run

  1. Open Settings, choose your vault path, and click Save.
  2. Click Rebuild index. The dashboard and search become available once the first index completes.
  3. Retrieval mode defaults to Keyword; no extra setup is needed.
  4. For semantic or hybrid search, enter an OpenAI-compatible or Ollama embedding endpoint, model, and dimensions in Settings. Click Test connection, save, then rebuild the index.
  5. The Allow vault context in conversations switch on the dashboard is off by default. Turn it on when you want the Agent to use the vault; when off, every obsidian_* tool call is denied.

Settings

Usage

Dashboard

The home page shows note, word, tag, 7-day update, and chunk counts; a 365-day heatmap; a 14-day activity curve; Agent call counters; recent notes; and tag distribution.

Search

Type a keyword or prefix on the Search page; results update as you type. Filter by folder or tag, add passages to the Context tray, then copy a source-attributed context packet or let the Agent search through tools.

Search

Notes

Browse and read notes on the Notes page. Model writes follow a proposal flow:

  1. The model proposes new or appended content.
  2. The page shows the target path and a diff.
  3. Nothing is written until you approve.
  4. Writes can be undone and are backed up.

Use in a conversation

  1. Turn on Allow vault context in conversations on the dashboard.
  2. Ask the model to search your notes, for example:
    • "Find my notes about RAG and cite the sources."
    • "Append today's meeting notes to the project note."
  3. Only passages actually retrieved or read enter the conversation; the whole vault is never injected automatically.
  4. Turn the access switch off when you no longer need it. Later tool calls will be denied.

Agent tools

ToolPurpose
obsidian_searchSearch note passages by mode, folder, or tag
obsidian_read_noteRead a full note
obsidian_list_notesList or filter notes
obsidian_statsRead index, statistics, and workspace status
obsidian_propose_writeCreate a create/append proposal without writing
obsidian_apply_writeApply a proposal after your approval
obsidian_undo_writeUndo a write from this session
obsidian_index_statusCheck indexing and embedding progress

The plugin has no delete, move, rename, or overwrite tools.

Privacy and safety

  • The index, vectors, settings, and usage counters are local. The plugin contains no telemetry and does not upload the vault.
  • When Allow vault context in conversations is off, every obsidian_* Agent tool is denied before execution. Vault Studio's own search and browsing remain available.
  • The system prompt tells the model to report Obsidian tool failures and never bypass Vault Studio or the access switch through generic file tools.
  • The access switch only gates obsidian_* tools. For a hard boundary, also check DeepSeek Harness workspace/sandbox permissions so generic read, grep, and bash cannot reach the vault path.
  • Writes require approval, keep backups, and can be undone. The plugin never deletes or overwrites notes automatically.

License

MIT

Comments

Loading…

Similar plugins

dsh-obsidian

by mingzeng21

Connect DeepSeek Harness (dsh) to a local Obsidian vault: search, read, write, move, and trash notes.

Memory & ContextManifest valid

★ 10

↓ 1k/wk

MIT

TypeScript

Aug 23, 2026

dsh plugin --profile web add dsh-obsidian

by zhaoxuejie

让 DeepSeek Harness agent 把本地 Obsidian 知识库当作长期记忆与工作台:全文/语义检索、会话记忆注入、一键捕获、巡检管家(只建议不擅改)

Development & InfrastructureManifest valid

★ 1

MIT

JavaScript

Sep 7, 2026

dsh plugin --profile web add dsh-plugin-vault-memory

by bufan528

Obsidian-compatible long-term memory vault plugin for DeepSeek Harness (DSH): local Markdown storage, sidebar management page, token-budgeted recall, durable journal outbox.

Security & AuditUI & ExperienceMemory & ContextSessions & MessagesManifest valid

★ 7

↓ 2.2k/wk

MIT

TypeScript

Oct 29, 2026

dsh plugin --profile web add dsh-longterm-memory

by chenkezhen480

Semantic long-term memory for DeepSeek Harness: embedding-based retrieval over a persistent cross-session memory store, with model-facing tools, proactive per-question recall, automatic conversation summarization, and auto-selected embedding provider (apiKey present → API, otherwise local).

Tools & CapabilitiesManifest valid

★ 0

Apache-2.0

TypeScript

Aug 17, 2026

dsh plugin --profile web add dsh-plugin-semantic-memory

by M1r0er

Turns a local Markdown vault into two folders the user picks in Settings: a knowledge folder DSH writes notes into, and a memory folder used as the agent's long-term memory with per-turn recall and en

Memory & ContextManifest valid

★ 0

JavaScript

Oct 24, 2026

dsh plugin --profile web add dsh-inkvault

by 173787247

DeepSeek Harness WSL bridge to a Windows Obsidian vault (status/open/read/write/search).

Tools & CapabilitiesDevelopment & InfrastructureManifest valid

★ 1

MIT

JavaScript

Nov 5, 2026

dsh plugin --profile web add dsh-wsl-obsidian